الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 170 |  |  |
| | | from | 170 | | |
Abstract
Today medical imaging technology provides the clinician with a number of complementary diagnostic tools such as x-ray computer tomography (CT), magnetic resonance imaging (MRI) and ultrasound. Magnetic resonance (MR) imaging has been widely applied in biological research and diagnostics, primarily because of its excellent soft tissue contrast, non-invasive character, high spatial resolution and easy slice selection at any orientation. In many applications, its segmentation plays an important role on the following sides: (a) identifying anatomical areas of interest for diagnosis, treatment, or surgery planning paradigms; (b) preprocessing for multimodality image registration; and (c) improved correlation of anatomical areas of interest with localized functional metrics. Brain is one of the most complex organs of human body so it is a difficult task to differentiate its various components and deeply analyze them. MRI are very common for brain image analysis. To analyze brain MRI data to bring useful information for diagnostic, MR images must be segmented into different tissues that composing it. The main brain tissues are white matter (WM), gray matter, and cerebrospinal fluid. Moreover, regional volume calculations of these tissues may bring more useful diagnostic information. For example, the quantization of gray and white matter volumes may be of major interest in neurodegenerative disorders such as Alzheimer disease, in movements disorders such as Parkinson or Parkinson related syndrome, in white matter metabolic or inflammatory disease, or perinatal brain damage, or in post traumatic syndrome. Previously, in many clinical studies segmentation is still mainly manual or strongly supervised by a human expert. The level of operator supervision impacts the performance of the segmentation method in terms of time consumption, leading to infeasible procedures for large datasets. Moreover, due to the characteristic of MR images, there are mainly three considerable difficult such as noise and intensity in-homogeneity that make the manually segmenting result are not accurate. The main contribution of the work we present here is a novel framework for fully automatic brain tissue segmentation. The proposed system consist of the following steps: First, remove of non-brain tissues, correct the intensity in-homogeneity, and finally segment the brain tissues. Therefore, the system will be able to accurately segment the brain tissues and calculate a regional volume of these tissues to bring useful information for diagnostic. The system consists of three-step segmentation procedures. First, non-brain structures removal by hybrid watershed Algorithm (HWA). Then, the accuracy of the intensity in-homogeneity correction may be improved through an iterative procedure of adaptive method, which is based on computing estimates of tissue intensity variation throughout the image. Finally, it uses a Fuzzy Kohonen Competitive Learning (F-KCL) for MRI brain image segmentation. The efficacy of the proposed system is demonstrated by extensive segmentation experiments using both simulated and real MR images. Extensive experiments show that the proposed framework can produce good segmentation results.